James McMurray said:
The difference between the methods is still always a single point.
Yes, this is true, but you miss the
point.
One point can have a significant difference for the D4+1 ability score spells. It has little difference for Magic Missile.
If you use the add each empower to the effect method, either a 1 or 2 roll (odd ability score) or 3 or 4 roll (even ability score) will result in no increase due to rolling one more on the roll.
If you use the multiplier method, it matters not what your ability score is. A 4 roll always does better than a 3 roll which always does better than a 2 roll which always does better than a 1 roll. Always (for Empower, Empower, Empower).
Hence, his statement that "It will matter more for the 1d4+1 spells." is true.
And, in fact, for Empower Empower Bull Strength, a similar problem occurs. The add up each empower to the effect method sucks for even ability score characters, but always increases the ability score modifier for odd ability scores:
1:4 (2 + 1 + 1)
2:5 (3 + 1 + 1)
3:8 (4 + 2 + 2)
4:9 (5 + 2 + 2)
The multiplier method is just as good for each:
1:4 (2 * 2)
2:6 (3 * 2)
3:8 (4 * 2)
4:10 (5 * 2)
You asked "How do you figure?", I told you how. The multiplier method is smoother, hence, it doesn't have weird jumps and discrepencies in it.